Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : LAST MILE TECHNOLOGIES - Let's Discuss Them Here -- Ignore unavailable to you. Want to Upgrade?

To: Frank A. Coluccio who wrote (3202)	3/22/1999 11:03:00 AM
From: John Stichnoth	Read Replies (2) \| Respond to of 12823

As usual, you've taken an issue and broken it down concisely into a number of sub-issues. First, to show my ignorance: In my post, I mentioned 4 steps on "Iconnet". But I don't know what that is. Is that Bell Atlantic's private pipe, or is it a public pipe that BA is using? All the tracerts I've run have shown iconnet in the path, so I assume it's a BA pipe. I really was making just two points in my discursion. First, that the number of jumps being made by the "last mile" vendor is too high, whether it's 4 or 8 hops to get onto the backbone. Maybe this is peculiar to BA. But, when you're making that many hops you are necessarily sitting around in too many caches. Which brings us to... Second, the time at each router/switch is taking too long. CSCO et al are becoming the bottleneck. Each router/switch has a cache, I presume. Whenever there is a packet sitting in one of those caches, that means that all the messages being sent through that router/switch are being slowed down--assuming they are being handled on a FIFO basis. Then all of the sudden the cache overflows and a packet is dropped entirely. (Note on a related issue: In terms of prioritizing messages, is there any work being done on attaching a priority code to messages, so "important" ones--ie., those for which there are "reason$" to make sure they get there faster--can be given priority?) The peering of networks (network of networks) does not seem, to me, to be where the problems will occur, either in terms of the number of steps or in the time through those steps. As needed, the pipeline maintainers have the expertise, wherewithal and motivation to keep up to date with their switches. They can go out and get the latest juniper switch. And they will drive the delay through their switches down to the 20 ms goal. (If they're sitting there with all of that investment in fibre, they won't want to see it sit empty just because their own on-ramps are clogged--to mix a metaphor). On your "almost free" point: The pipe is already "almost free", from the end-user's standpoint. It seems to me that the bulk of the costs are in the ISP, for billing, other customer support, installations (in the CO and on the customer's premises) and maintenance. Does anyone have an estimate on how much is being spent on backbone development versus local development? Private lines are not in issue, as far as I'm concerned. Maybe you can explain more about the onnet/inc/ion alternatives. I had thought of them as merely an extension of the network of networks issue. They may save a hop or two, but really are designed to keep business on each company's wire for as long as possible. "Hopping" efficiencies were secondary considerations at best. They only become more important to the extent they become an oligopoly and prevent second tier names from peering. (Which may be an issue in the future?). The problem comes down to the ISP. The Last Mile, and within the local ISP's infrastructure. There. I've said my piece. I'll be quiet for awhile. Thanks for listening. Best, JS